tensorflow 20 13-10 iter:0 ,test_acc:0.098 iter:10 ,test_acc:0.098 iter:20 ,test_acc:0.098

iter:0  ,test_acc:0.098
iter:10  ,test_acc:0.098
iter:20  ,test_acc:0.098
import tensorboard
import tensorflow as tf
import time
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/home/mj/MINIST_data", one_hot=True)

with tf.name_scope("input"):
    with tf.name_scope("x_input"):
        x=tf.placeholder(tf.float32,[None,784],name="x_input")
    with tf.name_scope("y_input"):
        y=tf.placeholder(tf.float32,[None,10],name="y_input")

#将图像reshape后，送入第一层卷积层   卷积核为5x5,将1通道图像卷积为6通道
with tf.name_scope("conv1"):
    x_image=tf.reshape(x,[-1,28,28,1])
    conv1_filter1=tf.Variable(tf.truncated_normal([5,5,1,6]))
    conv1_bias1=tf.Variable(tf.truncated_normal([6]))
    conv1=tf.nn.conv2d(x_image,conv1_filter1,strides=[1,1,1,1],padding="SAME")
    h_cv1=tf.nn.relu(conv1+conv1_bias1)
    h_cv1_maxpool=tf.nn.max_pool(h_cv1,ksize=[1,2,2,1],strides=[1,2,2,1],padding="SAME")

#送入第二层卷积层   卷积核为5x5,将6通道图像卷积为16通道
with tf.name_scope("conv2"):
    conv2_filter2=tf.Variable(tf.truncated_normal([5,5,6,16]))
    conv2_bias2=tf.Variable(tf.truncated_normal([16]))
    conv2=tf.nn.conv2d(h_cv1_maxpool,conv2_filter2,strides=[1,1,1,1],padding="SAME")
    h_cv2=tf.nn.relu(conv2+conv2_bias2)
    h_cv2_maxpool=tf.nn.max_pool(h_cv2,ksize=[1,2,2,1],strides=[1,2,2,1],padding="SAME")

# 送入第三层卷积层   卷积核为5x5,将16通道图像卷积为120通道
with tf.name_scope("conv3"):
    conv3_filter3 = tf.Variable(tf.truncated_normal([5, 5, 16, 120]))
    conv3_bias3 = tf.Variable(tf.truncated_normal([120]))
    conv3 = tf.nn.conv2d(h_cv2_maxpool, conv3_filter3, strides=[1, 1, 1, 1], padding="SAME")
    h_cv3 = tf.nn.relu(conv3 + conv3_bias3)


#第一个全连接层
with tf.name_scope("w_fc1"):
  w_fc1=tf.Variable(tf.truncated_normal([7*7*120,80]))
  b_fc1=tf.Variable(tf.truncated_normal([80]))
  f_flat=tf.reshape(h_cv3,[-1,7*7*120])
  h_fc1=tf.nn.relu(tf.matmul(f_flat,w_fc1)+b_fc1)

#第二个全连接层  输出层
with tf.name_scope("classfier"):
  w_fc2=tf.Variable(tf.truncated_normal([80,10]))
  b_fc2=tf.Variable(tf.truncated_normal([10]))
  h_fc2=tf.nn.relu(tf.matmul(h_fc1,w_fc2)+b_fc2)
#损失函数
with tf.name_scope("loss_function"):
  loss=tf.reduce_sum(y*tf.log(h_fc2))
  tf.summary.scalar("loss",loss)
#优化器
with tf.name_scope("optimizer"):
    train_step=tf.train.AdamOptimizer(0.01).minimize(loss)
#求精确du
with tf.name_scope("accuracy"):
    correct=tf.equal(tf.argmax(y,1),tf.argmax(h_fc2,1))
    accuracy=tf.reduce_mean(tf.cast(correct,tf.float32))
    tf.summary.scalar("accuracy",accuracy)
merged=tf.summary.merge_all()

#启动会话

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    train_writer=tf.summary.FileWriter("los1",sess.graph)
    for i in range(10001):
      batch_x,battch_y=mnist.train.next_batch(batch_size=200)
      sess.run(train_step,feed_dict={x:batch_x,y:battch_y})
      summary=sess.run(merged,feed_dict={x:batch_x,y:battch_y})
      train_writer.add_summary(summary,i)
      if i%10==0:
          test_acc=sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels})
          print("iter:"+str(i)," ,test_acc:"+str(test_acc))